Cardiac shunt in COPD as a cause of severe hypoxaemia: probably not so uncommon after all
- B.G. Boerrigter,
- A. Boonstra,
- N. Westerhof,
- P.E. Postmus and
- A. Vonk-Noordegraaf⇑
- A. Vonk-Noordegraaf, Department of Pulmonary Diseases, VU University Medical Center, De Boelelaan 1117, 1081 HV Amsterdam, The Netherlands. E-mail: a.vonk{at}vumc.nl
To the Editors:
Hypoxaemia is a common finding in chronic obstructive pulmonary disease (COPD) and may be aggravated during exercise. The main mechanism is perfusion through areas that are not well ventilated: a ventilation–perfusion mismatch. True shunting, defined as venous blood mixing directly with end-capillary blood at the arterial side of the circulation, is not a usual cause of hypoxaemia in COPD 1, 2. The amount of shunting related to ventilation–perfusion mismatch is usually made by the calculation of the shunt fraction while the patient is inhaling 100% oxygen. We present two cases of severe hypoxaemia in patients with COPD to show that a cardiac shunt can contribute to hypoxaemia, and that this shunt can be missed if the 100% oxygen method is used to quantify shunt.
Patient A was a 68-yr-old man, referred from another hospital to our clinic for evaluation of severe dyspnoea and hypoxaemia. Patient complaints were rapidly progressive dyspnoea and severe impaired exercise tolerance with blue discoloration of the fingers and lips for 3 months. There were no complaints of coughing, sputum production or fever. Patient A ceased smoking 12 yrs previously after 60 pack-yrs. Physical examination revealed peripheral and central cyanosis, and a little ankle oedema. Arterial blood gas analysis revealed a hypoxaemia at rest, with an arterial oxygen tension (Pa,O2) of 42 mmHg and an arterial oxygen saturation (Sa,O2) of 77%. Pulmonary function tests showed moderate obstruction (forced expiratory volume in 1 s (FEV1) of 67% predicted and FEV1/synchronised vital capacity (SVC) ratio of 65% pred). Initiation of exercise resulted in immediate oxygen desaturation to 71%, preventing any further exercise. Radiological analysis by means of high-resolution computed tomography (CT), pulmonary angiography and ventilation–perfusion scan revealed only signs of emphysema. Transthoracic and transoesophageal echocardiography showed dilation of the right atrium and ventricle with signs of pulmonary hypertension. No signs of an intracardiac right-to-left shunt were seen. Both tests were performed while 15 L·min−1 oxygen was administered. Right heart catheterisation confirmed the diagnosis of pulmonary hypertension (table 1). The response to administration of 100% oxygen was a fall of mean pulmonary arterial pressure (Ppa) to 24 mmHg, accompanied by a significant increase in arterial oxygenation to 96%.
For this reason, we assumed that a right-to-left shunt through a patent foramen ovale (PFO) was present under normoxic conditions and during exercise, although no intracardiac shunt was seen during echocardiography. Cardiac magnetic resonance imaging (MRI) was performed while the patient was breathing room air. Flow per beat was measured in the pulmonary artery (44 mL) and aorta (56 mL), which resembled a right-to-left shunt of 25%. We concluded from these measurements that oxygen desaturation during exercise under normoxic conditions occured as a consequence of right-to-left shunting through a PFO. The patient was referred to a cardiologist for a percutaneous closure of the PFO, which resulted in an improved clinical condition together with an oxygen saturation of 95% at rest which remained unaltered during exercise. Control MRI showed similar flow per beat through the aorta and pulmonary artery (57 mL).
Patient B was a 70-yr-old male with a history of COPD, hypercapnia and echocardiographic signs of pulmonary hypertension (PH). Patient B complained of progressive dyspnoea on exercise with blue discoloration of lips and fingers for several months. He had experienced no acute exacerbations of his COPD in the previous year. Patient B ceased smoking 18 yrs previously. Arterial blood gas analysis showed hypoxaemia (Pa,O2 48 mmHg; Sa,O2 81%) and hypercapnia (carbon dioxide tension 56 mmHg). Hypercapnia was interpreted as a sign of respiratory insufficiency for which a trial of bilevel positive airway pressure (BiPAP) treatment was tried in the past, resulting in a further worsening of his hypercapnia and hypoxaemia. Pulmonary function tests showed severe obstruction (FEV1 41% pred and FEV1/SVC 37% pred). During cardiopulmonary exercise testing, patient B reached 70 W (44% pred), which led to an oxygen desaturation to 71% while heart rate reserve was 21% and ventilatory reserve was 50%. For this reason, hypoxaemia was considered the exercise-limiting factor. High-resolution CT, pulmonary angiography and ventilation–perfusion scan showed severe emphysema without evidence for pulmonary embolisms. Echocardiography, with oxygen (15 L·min−1), showed dilatation of right atrium and right ventricle, and signs of PH. No intracardiac shunt was found. Right heart catheterisation was performed, which revealed increased Ppa and right atrial pressure (RAP) (table 1). Inhalation of 20 ppm nitric oxide led to a normalisation of Ppa and oxygen saturation. Based on this finding, it was concluded that the severe desaturation during exercise was most likely due to right-to-left shunting through an PFO in the presence of PH secondary to COPD. Echocardiography was repeated under normoxic conditions, showing a PFO at a location unsuitable for percutaneous closure. Because of the good response to inhaled NO a trial of a phosphodiesterase (PDE)-5 inhibitor was initiated, leading to clinical improvement and a stable oxygen saturation of 90%. After 4 yrs, patient B was re-evaluated, showing a stable condition with a Sa,O2 of 91% at rest with a similar flow per beat through the aorta and pulmonary artery (70 mL) measured with MRI.
DISCUSSION
A PFO is present in 25–30% of the population; however, most will remain without haemodynamic consequences under normal conditions 3. In case of increased RAP, right-to-left shunting through a PFO might occur, leading to hypoxaemia 4, 5. We presented two cases of right-to-left shunting through a PFO in the presence of PH secondary to COPD, in which shunting significantly contributed to the hypoxaemia. The PH presumably started shunting through the PFO, leading to hypoxaemia. With hypoxaemia, the oxygen pressure in mixed venous blood also decreased, thereby stimulating hypoxic pulmonary vasoconstriction 6, in turn leading to an increase in Ppa. This led to a further increase in RAP and, thereby, to a larger pressure gradient over the PFO, leading to the vicious circle, as shown in figure 1. This mechanism also explains the severe desaturation observed at exercise, as cardiac output augmentation induced by exercise will lead to an increase in Ppa and, thus, an increase in the right-to-left shunt. Finally, BiPAP might induce an increase of the right-to-left shunt, as was observed in patient B, due to its effect on Ppa 7.
In both patients, active vasoconstriction contributed to PH to a great extent, as was proven by the effects of the vasodilating agent nitric oxide. Arterial oxygenation improved with vasodilation, in contrast with large groups of COPD patients 8, 9. This provides evidence that the right-to-left shunt was the main mechanism of hypoxaemia in our patients. Therefore, it was decided to close the PFO in patient A as a treatment of the cause of the PH, as outlined in figure 1.
Because of the inability to close the PFO percutaneously in patient B, a trial with a PDE-5 inhibitor was initiated. Although studies revealed no benefit of PDE-5 inhibitors or, even, worsening of the ventilation–perfusion match in COPD patients 9, 10, we started it based on the beneficial response to inhaled NO. Our aim was to lower Ppa and RAP, leading to closure of the PFO. Because of a good response and clinical improvement, we decided to continue treatment with the PDE-5 inhibitor. However, no conclusions should be drawn from our observations on how we should treat COPD patients with PH in general. Both patients were already on oxygen treatment, which was continued afterwards, since oxygen therapy is the only proven beneficial therapy for COPD and PH.
Why was initially the PFO not seen with echocardiography in both patients? Most probably because these measurements were performed at rest with administration of 15 L·min−1 oxygen. With the vasodilatory effect of oxygen, RAP, the driving force of the right-to-left shunt, was lowered and the PFO was physiologically closed.
Thus, where pulmonary shunting in patients with COPD is usually not a cause of hypoxaemia, right-to-left shunting through a PFO can contribute to the hypoxaemia in presence of secondary PH. The importance of the shunting might be underestimated if echocardiographic evaluation is performed while the patient uses high-fraction oxygen therapy. Traditionally, it is assumed that a trial with a high inspiratory oxygen fraction is a good measure to quantify shunt. Our two cases showed that, in the case of COPD, this rule does not always apply.
Footnotes
Statement of Interest
A statement of interest for A. Boonstra can be found at www.erj.ersjournals.com/site/misc/statements.xhtml
- ©2011 ERS